In many modern digital image systems image frames are buffered for further image processing. In order to reduce the required memory size as well as required I/O bandwidth, compression of the buffered images may be employed. The requirements of the compression algorithm for this kind of application usually include low complexity, fixed compression ratio and high quality (e.g., visually lossless). On the other hand, the required compression ratio is usually allowed to be lower, e.g., 2 to 4 would be sufficient.